基于网格搜索与支持向量机的轴承故障诊断

A Bearing Fault Diagnosis Method Based on Grid Search and Support Vector Machine

针对轴承故障诊断问题,提出一种基于相关度分析与网格搜索算法(GS)优化支持向量机(SVM)的轴承故障诊断方法。采用GS算法对SVM的惩罚参数c和核函数参数g进行寻优,以此建立分类器用于识别轴承故障类型。在模型建立方面巧妙地加入了分层的思想,通过相关度分析之后采用多层GS-SVM模型使轴承的故障诊断准确率相对于近年来的研究得到了明显的提升。最后,采用凯斯西储大学轴承数据中心的滚动轴承故障数据进行了分类识别实验。实验表明,研究提出的轴承故障诊断方法在直接作用于原信号的基础上不仅能够有效的识别轴承正常状态、内圈故障、外圈故障以及滚珠故障,而且还对每一类故障的严重程度有很好的区分,提高了故障类样本的诊断正确率,具有较强的实用性。

Aiming at the problem of bearing fault diagnosis,a method of bearing fault diagnosis based on correlation analysis and support vector machine(SVM)optimized by grid search optimization(GS)was proposed.Grid Search algorithm was used to optimize SVM's penalty parameter and kernel function parameter,so as to establish a classifier for bearing fault identification.In the aspect of model establishment,the idea of stratification was subtly added.After the correlation analysis,the multi-layer GS-SVM was adopted,which significantly improved the accuracy of bearing fault diagnosis compared with recent researches.Finally,the rolling bearing fault samples from case western reserve university were used in the classification and identification experiments.The experimental results show that the proposed bearing fault diagnosis method can not only effectively identify the normal condition of bearing,inner ring fault,outer ring fault and ball fault,but also distinguish the severity of each kind of fault,improve the diagnostic accuracy of fault samples,and has strong practicability.